SPH Master class

[Hotel26]

I'm working on my second spaceplane and I'm stuck.  This is the craft file for Aquila Falcon.

I think that any real master, looking at my work, is going to know pretty fast what I don't know and be able to offer a critique that will benefit not only me but maybe a few other noobs.

I am following three requirements:

1. It's a surface-to-orbit shuttle for duty on Kerbin and Lathe, so it must have a Mk-3 cabin
2. I want to be able to couple it to an interplanetary vehicle (as well as launch it from Kerbin that way when I feel too lazy to wrestle it), so it needs a Senior docking ring for a vertical launch
3. In order to fit in with the rest of my fleet, it has to be butt-ugly.

Falcon, as it is, has the following problems that I am aware of and many more, I'm sure:

• the landing gear aren't straight, but cocked slightly forward
• I don't understand what the SPH is telling me about the Center of Lift, but I think that is the key to:
• a lack of pitch authority, hampering landings.  I've landed it but only on the back of the power curve...
• the tail fins are unreasonably butt-ugly and are producing drag via downforce.
• maybe the wings are on upside down -- I can't tell; it's one explanation for the crazy lift vector.
• I can't land it heavy with fuel -- it breaks and explodes
• the MP tank is unreasonably butt-ugly but I don't know anything about cargo holds either

(Sorry, no screenshots as I am actually without internet at home at the moment.)

I'm hoping that one or more Grand Masters will comb over this file and not only a) list out every last little flaw and nit with this craft, but also b) teach the technique in the SPH for everything I've been unable to do.

Thanks in advance!  Go, team.

Edited August 16, 2016 by Hotel26
fixed link

[bewing]

1. Your craft file says I need a decryption key to get the DL.

2. The way to get your wheels straight is to use Absolute Rotation Mode. Click the rotate gizmo. Make sure you are in Angle Snap mode. Click a wheel. Hit the F key. Now you are in Absolute Rotation Mode. It's a little harder to see than normal rotation mode, but if you rotate something so that it looks straight, it is absolutely straight.

3. Wings are ambidextrous. There is no upside down.

4. So don't land it heavy with fuel.

5. You can use the offset tool and clip the MP tank inside some other part? But cargo bays are quite simple. They give you two attachment nodes inside the bay for building stacks of little stuff, and you can attach radial mount stuff anywhere inside.

6. You can use smaller crew pods inside a cargo bay .... How many kerbals do you really need to lift at one time into LKO? I never need more than 9. So isn't the MK3 deal overkill?

[Snark]

I'd also strongly suggest posting screenshots.  They're actually more helpful than a .craft file, for lots of folks.

[Hotel26]

I fixed the link in the original post: please pardon me.

This is an update, though.  I completely rebuilt the craft, keeping an eye on the center-of-lift indicator the whole time.  It looks like it was some kind of self-inflicted GSW.

The difference is marked, with the new version able to climb from sea level with about 35° attitude.

@bewing, thanks for the helpful comments.  ()I definitely need an Mk-3 for colonization purposes!)
 

 

Edited August 16, 2016 by Hotel26
photos inline

[Snark]

Took the liberty of fixing your photo links for you. 

From the picture, it's pretty clear why you have a lack of pitch authority.  Look at your second picture.  In particular, note where the CoM is... now look where most of your control surfaces are (on the trailing edge of the wing). Really, really close, right?  (Doesn't matter how far out they are to left and right-- I'm talking about distance along the ship's longitudinal axis).  The fact that they're so close to the CoM means that they have a really short lever arm to work with, which means they have very little control authority.

The canards are well-placed (nice long distance in front of CoM), but they're just not big enough to make all that much of a difference to that big heavy plane.

Suggestion:

• Replace the canards with a pair of wings, and put the biggest ailerons you can on them. That'll give a lot more pitch authority.
• You've got a couple of stubby little horizontal elevators on the tail.  Make them bigger and move them as far back as you can, ideally right next to that docking port on the rear.
• Make sure you've turned off roll authority on your tail fin.  This will help with yaw stability, though it won't make much difference to pitch authority.
• If your landing gear are exploding, use bigger gear.  If you're already using the biggest gear, use more of them.

[GoSlash27]

I concur with what Snark said.

Adding to this, you can do with less intake area and more static incidence. Also, those airliner wings are prone to overheating on reentry. I recommend the shuttle wings instead.

Best,
-Slashy

[foamyesque]

You're carrying more intakes than you need. One shock intake will feed six rapiers no problem; nose-mount it on the pod and use nosecones to lead the rapier stacks.

 

I disagree with @Snark; the two canards you've got should be able in themselves to provide pitch authority.

EDIT: Wait, those aren't standard canards, those are delta-delux winglets. Never mind! @Snark is right, you need to upgrade those. Move them to standard canards.

I would suggest pulling both them and your wing forward, to bring your CoL closer to, or even inside, your CoM. That will dramatically amplify the strength of your canard's controls. The giant pile of control surfaces on the trailing edge of the wing can be trimmed to just a pair of ailerons for roll control, and maybe another pair to act as flaps. Likewise, those tiny h-stabs at the back can be deleted. Your tailfin is awfully large; I would downgrade it to a standard canard.

 

@GoSlash27 has a point about using the Big-S wings instead of the FAT ones; the FAT ones have a much lower heat tolerance. You can manage with them, but there's less room for error.

Edited August 16, 2016 by foamyesque

[Hotel26]

Gentlemen, congratulations, Falcon 5 is in orbit.  The 4 test pilots hated it at first sight, saying it was the butt-ugliest space vehicle they had ever beheld, but also the wildest ride to space any had ever endured.  There's not an unused barf bag to be found aboard.

Seriously, guys, excellent work!!

What makes it particularly sinister are the S-wings, packed with Rapiers.  I switched wings after a couple of test flights exploded around 45 km but later realized that was MJ fooling with physics warp.  Serves me right for relinquishing the controls on a test flight.

Photos tomorrow.

Oh yeah.  I haven't tried a landing yet.

[foamyesque]

Here's my shot at a redesign. I didn't realize you were using the DD winglets instead of standard canards, @Hotel26; standard canards, if you've got access to them, are my go-to control surface. They provide great gobs of control authority. It also turns out that with the very short lever arc I needed a bigger tail than I'd planned on to maintain yaw stability.

 

Either way:

 

https://www.dropbox.com/s/wi04ugmvg0iz4bb/SPH Master revision.craft?dl=0

 

Can land at <60m/s even with full fuel load. Carries a bit less fuel (~400 units) than yours; the Rapiers are mounted on structural fuselages to balance the CoM and cut down on its movement.

Edited August 16, 2016 by foamyesque

[Hotel26]

@foamyesque, you are a champion.  I can't wait to test fly your plane!  (Then my minions are going to pull it apart to learn all its secrets!)  I'll report back here when we're finished.  I also have some photos of Falcon 5 to post but -- apologies for the slow turn-around -- no internet at home for a week, awaiting an upgrade...

I'll also post here some screenshots of your creation.  (Ugh, I see you did post a screenshot...  my work internet censor ate it.  )

(re: canards, I think I had them on the very original Falcon that had the CoL problems.  I made that one worse with shuttle fins on the rear which is why the CoL wound up being back in the tailplane pushing upward and forward.  I didn't realize that canards do act as controls because they're not animated; so that is very good information: thank you.)

Edited August 17, 2016 by Hotel26
censorship

[dire]

I may be missing something but I don't see the ample quantities of RCS thrusters you're going to need in order to dock this beastie backwards. It could be that you intend this craft to sit idly by while something else docks to it but if you are planning a colonization mothership (which is what it sort of sounds like) then this will be your "small" craft and should do the grunt work of the docking.

You probably also want to add a Mk 2 RCS tank to it somewhere. I've docked shuttles like that before and they eat up a ton of monoprop with every dock. I also like to put my docking ports on the nose (which makes it much much easier) but that might not be feasable with this design.

Edit: I see the RCS thrusters on the nose of the original craft but not the improved model. You also want to distribute them evenly around center of mass to minimize the work your SAS will do while docking, for example six forward and six aft.

Edit Edit: You can estimate your RCS dV with Mechjeb and a "Puff" monoprop engine. I would shoot for about 120 dV in monoprop per docking attempt you want to do. If you are not good with docking, maybe take 200 dV and quicksave beforehand. (Me personally, I like using the main engines until I get to within 100 meters, and then using just the barest whiffs of monoprop, but I also mount my grapplers on the nose -- I don't use docks as such. I'd still bring at least 60 dV).

Edited August 17, 2016 by dire

[Spricigo]

1 hour ago, dire said:

I may be missing something but I don't see the ample quantities of RCS thrusters you're going to need in order to dock this beastie backwards. 

And I'm pretty far from any point but I'm curious.  Why backwards would make a difference in space?  Even more since when you select control from here in the docking port it will be forward. 

I understand that it can be confusing because apparent discrepancy between control imput and ship movement. Just curious about how much an issue it can be or if I'm missing something 

 

[Hotel26]

The minions report:

78 parts.
 mk3Cockpit.Shuttle_4294710386
   mk3CrewCabin_4294049204
     mk3FuselageLFO.50_4292087608
       adapterMk3-Size2_4292087572
         asasmodule1-2_4292087536
           dockingPortLarge_4292087508
           linearRcs_4292083126
           linearRcs_4292083086
           linearRcs_4292083046
           linearRcs_4292083006
         structuralWing_4292083668
           elevon2_4292083634
       pointyNoseConeB_4292087474
         MK1Fuselage_4292087402
           Mk1FuselageStructural_4292087368
             Mk1FuselageStructural_4292087272
               Mk1FuselageStructural_4292086950
                 RAPIER_4292086758
                 strutConnector_4292083540
       pointyNoseConeB_4292087438
         MK1Fuselage_4292087176
           Mk1FuselageStructural_4292087142
             Mk1FuselageStructural_4292087046
               Mk1FuselageStructural_4292086854
                 RAPIER_4292086618
                 strutConnector_4292083602
       pointyNoseConeB_4292086478
         MK1Fuselage_4292086442
           Mk1FuselageStructural_4292086408
             Mk1FuselageStructural_4292086312
               Mk1FuselageStructural_4292086216
                 RAPIER_4292086120
               strutConnector_4292083416
       pointyNoseConeB_4292085980
         MK1Fuselage_4292085944
           Mk1FuselageStructural_4292085910
             Mk1FuselageStructural_4292085814
               Mk1FuselageStructural_4292085718
                 RAPIER_4292085622
               strutConnector_4292083478
       wingShuttleDelta_4288246692
         wingShuttleRudder_4288191604
         wingShuttleElevon1_4288183868
       wingShuttleDelta_4288246656
         wingShuttleRudder_4288192618
         wingShuttleElevon1_4288184250
       solarPanels5_4292082806
       solarPanels5_4292082778
       solarPanels5_4292082638
       solarPanels5_4292082610
     wingShuttleStrake_4292085410
       pointyNoseConeA_4292084892
         MK1Fuselage_4292084856
           Mk1FuselageStructural_4292084822
             Mk1FuselageStructural_4292084726
               RAPIER_4292084630
               strutConnector_4292083292
           GearMedium_4292084490
     wingShuttleStrake_4292085374
       pointyNoseConeA_4292085294
         MK1Fuselage_4292085258
           Mk1FuselageStructural_4292085224
             Mk1FuselageStructural_4292085128
               RAPIER_4292085032
               strutConnector_4292083354
           GearMedium_4292084244
     GearSmall_4292083998
     linearRcs_4292082966
     linearRcs_4292082926
     linearRcs_4292082886
     linearRcs_4292082846
     solarPanels5_4292082750
     solarPanels5_4292082722
     solarPanels5_4292082694
     solarPanels5_4292082666
   shockConeIntake_4293845280
   CanardController_4293784780
   CanardController_4293784556
 

[Hotel26]

3 hours ago, dire said:

I may be missing something but I don't see the ample quantities of RCS thrusters you're going to need in order to dock this beastie backwards. It could be that you intend this craft to sit idly by while something else docks to it but if you are planning a colonization mothership (which is what it sort of sounds like) then this will be your "small" craft and should do the grunt work of the docking.

You probably also want to add a Mk 2 RCS tank to it somewhere. I've docked shuttles like that before and they eat up a ton of monoprop with every dock. I also like to put my docking ports on the nose (which makes it much much easier) but that might not be feasable with this design.

Edit: I see the RCS thrusters on the nose of the original craft but not the improved model. You also want to distribute them evenly around center of mass to minimize the work your SAS will do while docking, for example six forward and six aft.

Edit Edit: You can estimate your RCS dV with Mechjeb and a "Puff" monoprop engine. I would shoot for about 120 dV in monoprop per docking attempt you want to do. If you are not good with docking, maybe take 200 dV and quicksave beforehand. (Me personally, I like using the main engines until I get to within 100 meters, and then using just the barest whiffs of monoprop, but I also mount my grapplers on the nose -- I don't use docks as such. I'd still bring at least 60 dV).

Sharp eyes.  You could be talking about either my screenshot or foamyesque, but foamyesque was working from my initial (pictured), so I'll take responsibility.

The original Falcon had 4 on the front and 4 on back.  I have a feeling I've not corrected the oversight on the rebuilt Falcon pictured, so thank you.  I haven't attempted a dock with it but I would have noticed.  It's got a reaction wheel which is enough for rotational adjustment and conserves MP, (especially with MJ driving!!).  Nevertheless, RCS should be symmetric, as you say.  In any case, the original Falcon had 8 and a 750-unit MP tank and it was enough to dock.  You'll notice in the picture above that I've deleted the ugly yellow MP tank and slung a 150-unit canister under the belly.  The MK3 cockpit has 100 for a total of 250 which is plenty for docking.

Thanks for picking this up and kudos to everyone who has contributed to this topic so far.  Bravo.

Edited August 17, 2016 by Hotel26

[Hotel26]

2 hours ago, Spricigo said:

And I'm pretty far from any point but I'm curious.  Why backwards would make a difference in space?  Even more since when you select control from here in the docking port it will be forward. 

I understand that it can be confusing because apparent discrepancy between control imput and ship movement. Just curious about how much an issue it can be or if I'm missing something 

 

I think dire is indicating that there is plenty of forward thrust available from the Rapiers.  But something else is needed to power the reverse direction and, yes, that is conventionally RCS, of course.

@dire, I do the same as you: main engines to the vicinity (<= 100m) and then RCS for the dock.

Edited August 17, 2016 by Hotel26

[dire]

2 hours ago, Spricigo said:

And I'm pretty far from any point but I'm curious.  Why backwards would make a difference in space?  Even more since when you select control from here in the docking port it will be forward. 

I understand that it can be confusing because apparent discrepancy between control imput and ship movement. Just curious about how much an issue it can be or if I'm missing something 

 

Backwards is more an issue because it means your main thrusters are pointed the wrong to help with minor docking course corrections. It's not much of an issue with a sufficiently nimble ship, but I'm not convinced that this ship has enough torque and power generation to count as nimble -- if you have to burn 45 degrees off retrograde for 6 m/s, and then burn at -135 off retrograde for 2 m/s (for example) this ship might be able to power it and it might not. If you are like me ... I think 50 km and 50 m/s is a pretty good value of "close enough" for matching orbits at a gross scale. From there I will correct course to set prograde onto the pro-target reticule at a comfortable speed (20-50 m/s is fine) and correct for orbital drift as I go.

Once I close to 10 km I drop to 20 m/s. At this point, if my docking port is on the rear of the craft, my main engines are done; I might as well offline them so I don't mess my approach up by accident; if that's the case I might as well slow to 10 m/s while I can, even though it means a larger orbital drift on final approach. In that case, everything from here on is RCS only.

 Once I get to 5 km I drop my speed to half and correct for orbital drift as I go. When I get about 1 km from target I'll drop all the way down to 2-3 m/s (if it's a smaller, more nimble craft I can hold off a little longer but I really really want to be lined up by 100-150 meters out). Then I time warp, do final fine adjustments and correct for orbital drift at 50 meters and (for a larger craft like this one) slow down to 1 m/s or less. Then time warp to 10 meters and do it again. 

All told, this typically takes about 120 delta-V once my orbit is pretty well "matched". I can use my main engines for at least half of that; if my docking port is on the front of my craft that number is closer to 90% as I can do all the orbital corrections using main engines instead of the less-efficient RCS thrusters and save my monoprop for just the tiny translation corrections at the end.

@Hotel26 thanks for the Kudos and remember your craft should be able to accelerate in the direction its port is facing; eight linear RCS thrusters may or may not be enough depending on if they are facing the right way or not. You can probably get away with pointing your nose at the target, accelerating, and then flipping, but I've tried that and even on a tiny craft with huge torque (think, just command pod, basically) it's a painful way to dock.

Another (kind of random) suggestion since you have the big reaction wheel, if you swap out your many small solar panels for a few larger ones you'll cut down your part count and increase your power generation.

Good luck and fast travels!

Edited August 17, 2016 by dire

[Hotel26]

My own upgrade to Falcon 5 implementing suggestions:

I did go back to Falcon 4, with the FAT 455 wings and attempt to refit it, too, but it couldn't manage rotation below 100 m/s on the runway, and had lots of control authority problems at high altitude/high attitude.

 

I'm pretty darn impressed with foamyesque's suggestion/idea of Big-S fins fitted to the ends of Big-S wings...!

[dire]

I like to keep my fins behind the wings both to decrease the drag profile (I'm hoping, anyway) and to move them further behind the center of mass. The farther back your elevons are, the more leverage they have to keep your nose pointed where you want it.

Here is an (admittedly horrible) example: 

Note the wings are designed so that COM generally coincides with COL, but the elevons are about as far back as I could put them. In theory I'm avoiding a little bit of a drag penalty by having the tail fins be part of the larger wing but that's more than offset by the radial nukes up front.

[foamyesque]

8 hours ago, Hotel26 said:

I'm pretty darn impressed with foamyesque's suggestion/idea of Big-S fins fitted to the ends of Big-S wings...!

 

Thanks.

 

I really would suggest changing out those delta-deluxe canards for the standard canards, though. With deltas, only those tiny strips at the back are actually control surfaces; they're mostly useful as, e.g., vertical stabilizers on small aircraft, where you don't want a lot of control authority but do want large amounts of passive stabilization in that axis. That's not the case with your pitch problems. In contrast, the full surface of the standard canard moves, which means if you give it a control input it applies much more torque to your ship. It will make your atmospheric flight much simpler.

The other thing you're going to have is that with all of your fixed weight midbody or before (engines, passenger module, pod), as you burn fuel your CoM is going to shift way, way forward. That'll try to lock you prograde, which will further impede your pitch control and make the ship want to lawndart. That's why I put the engines as far back as I could; they act as a counterweight to the passenger module and pod that way, dramatically cutting down on the CoM shift (as long as your fuel storage stays where it is-- hence the structural fuselages).

Edited August 17, 2016 by foamyesque

[Hotel26]

6 hours ago, foamyesque said:

 

Thanks.

 

I really would suggest changing out those delta-deluxe canards for the standard canards, though.

The other thing you're going to have is that with all of your fixed weight midbody or before (engines, passenger module, pod), as you burn fuel your CoM is going to shift way, way forward

I had hoped to post a pirep (pilot report) on foamyesque's machine but I am apparently still running 1.0.5 and can't upgrade (to at least 1.1.13) until internet is restored.  Oh well.

So I tried updating Falcon with a number of ideas exhibited by foamyesque.  I'm using 5 standard canards: two horizontally on the cockpit and three in the empennage (tailplane).  Beautiful.  I tried the S-fin wing extensions but no luck for me, yet.  (I'll revisit this after studying the foam machine in the air.)

Understand your second suggestion and will see how your machine flies.  In my case, flying SSTO is an interesting proposition because it has multiple phases.  Similarly, I do like an aircraft that not only responds to my command but also requires me to understand it as well, particularly if it is to deliver peak performance.  I seem to have hit a sweet spot with Falcon 5b now: the CoL is close enough behind the Com to provide maneuverability but not so close as to produce overshoot in pitch adjustments (due to rotational momentum).  What I like about the amount of fuel I am carrying -- over the CoM and in the tail -- is that I can dynamically adjust the CoM by pumping fuel fore and aft.  What this does is provides trim to give the canards full range of authority.  I pump it back for rotation on the runway and for steep climbs and I pump it forward for the 12km and 20km accelerations.  Having the canards in a relatively neutral position gives me that control range.  I realize that a lot of pilots (most in fact) prefer docile craft with no complications (and no sane pilot likes any treacherous craft), but I don't mind flying mildly interesting procedures.

My Kerbals have dubbed Falcon "the flying hotdog" now (and sometimes "pig-on-a-stick"), due to its ugly appearance.  It has grown on me, though.   I look forward to 1.1.13, new parts and test flying foamyesque's machine.

Edited August 18, 2016 by Hotel26
clarification